1. Introduction Congress’ inability to pass legislation is a defining characteristic of the modern era. Since the 1950s, legislative gridlock has doubled in frequency (Binder, 2014) and, in October 2013 alone, was estimated to have cost the U.S. economy $24 billion (Lowrey et al., 2013). It is not surprising that divisive policies, which benefit one party at the expense of the other, become gridlocked. However, policies that have strong bipartisan support can also fail to become law. For example, the Development, Relief, and Education for Alien Minors (DREAM) Act creates a pathway to citizenship for undocumented immigrants that entered the U.S. as minors. The first version of the bill was introduced in April 2001 by a Democratic Republican pair of senators and has ‘‘enjoyed bipartisan support for two decades” (Munoz, 2021).1 Yet, repeated proposals of the DREAM Act (sometimes bundled with other legislation) have continually failed to be enacted (Petts, 2021). The precise reasons for these failures are unclear and often debated. Some political commentators speculate that the proposal would pass under a conscience vote,2 suggesting that party leadership may be to blame.3 Others argue that legislators have blocked the proposal in the hope of using it to pass larger—and likely more divisive—reforms.4 Consistent with both of these views, Senate minority leader Mitch McConnell elaborated on the Republican party’s recent opposition to the Deferred Action for Childhood Arrivals (DACA) policy, a weaker version of the DREAM Act, ‘‘everybody is sympathetic with the DACA issue [but] I can’t imagine that we would take up an immigration-related bill, no matter how worthy it might be ...without insistence on our part that we address the obvious crisis at the border” — Carney (2021). In this paper, I explore the possibility that Congress’ ability to engage in policy bundling, whereby legislators bundle diverse and (possibly) unrelated proposals into a single piece of legislation, can result in some bipartisan proposals being gridlocked.5,6 This idea runs contrary to the prevailing wisdom that policy bundling reduces gridlock by (i) facilitating compromises and (ii) allowing legislators to more fully exploit their proposal (or agenda-setting) power (see, e.g., Buchanan and Tullock, 1962; Câmara and Eguia, 2017; Krutz, 2000, 2001; Sinclair, 1997). As stated by former U.S. Senator Lieberman (2020), ‘‘leadership often resorts to a tactic of passing unpopular ideas by attaching them to bills supported on both sides of the aisle.” In a dynamic environment, where alternatives to the status quo arrive over time, policy bundling also creates an incentive for parties to gridlock bipartisan alternatives. By blocking a bipartisan proposal today, a party may benefit tomorrow since they can bundle this same proposal with a divisive alternative that otherwise would not pass. In other words, the prospect of policy bundling turns bipartisan proposals into a form of ‘‘leverage” that may be better used in—or saved for—future negotiations. Per Senator Lieberman (2020), ‘‘Some of the most popular ideas are in many cases held in abeyance so that they can be used to pass legislation that most members would oppose if considered separately.” Although it is difficult to say with certainty when and why bipartisan policies fail to pass, Senator Lieberman’s quote is a striking admission that policy bundling can have such an effect. It is also notable that Lieberman does not make reference to other potential causes of gridlock that are commonly proposed in the literature. For example, references to attempts to signal information to voters (Groseclose and McCarty, 2001; Patty, 2016) or an opposition (McCarty, 1997; Abreu and Gul, 2000), time constraints enforced by the legislative calendar (Chen and Eraslan, 2017), the difficulty in finding bipartisan policies (Acharya and Ortner, 2018; Callander, 2011), and concerns about an endogenous and hard-to-repeal status quo (Austin-Smith et al., 2019; Dziuda and Loeper, 2016, 2018) are all absent in Lieberman’s depiction. Instead, his depiction captures a simple—almost mechanical— mechanism that arises from policy bundling and can lead bipartisan policies to be gridlocked. Beyond simple anecdotes, there are a number of empirical observations that suggest that policy bundling can contribute to gridlock. Firstly—and in contrast to the U.S. Congress—in almost all U.S. state legislatures the possibility to policy bundle is restricted because of the single-subject and line-item veto rules (Brown, 2012).7 Yet, even in highly-polarized states, gridlock is less frequent than in the U.S. Congress (Masket, 2019). Furthermore, state legislators often cite the single-subject rule—and its effect of limiting policy bundling—as a key factor that helps them overcome gridlock (Center for Legislative Strengthening, 2018). This directly conflicts with the prevailing wisdom that policy bundling reduces gridlock and thus far the existing literature does not provide an explanation as to why such a pattern might arise. Secondly, Hughes and Carlson (2015) observe that salient bills in Congress that are (eventually) included in a policy bundle face substantial delays compared to stand alone legislation.8 In this paper, I study a dynamic model of bargaining that formalizes Senator Lieberman’s intuition. I establish that policy bundling can generate gridlock in equilibrium and then derive conditions that cause gridlock to be more likely to occur and study the implications for voter welfare. I also explore the model’s insights for the long-running debate of whether policy bundling should be removed from Congress via constitutional reform. I conclude with a discussion of the empirical implications of the results. I develop a benchmark model with two periods, two parties, and a voter. At the start of period 1, an alternative to the status quo is realized which may or may not be available in period 2, and the proposer decides whether to propose this alternative as a take-it-or-leave-it offer to the other party. At the end of period 1, the voter decides which party will be the proposer for period 2. In period 2, a new alternative to the status quo is realized, and the period-2 proposer decides whether to make a proposal. However, the set of possible proposals depends on the period-1 bargaining outcome. If the period-1 alternative passed in period 1, it will not be available as a proposal in period 2; otherwise, with some probability, it will be available as a proposal.9 In this latter case, the proposer decides whether to make a proposal and, if proposing both alternatives, decides whether to propose each alternative separately or to ‘‘bundle” the alternatives into a single proposal. By bundling the alternatives, the proposer forces the other party to either accept both alternatives or reject both. Each period’s alternative has a known voter-welfare component and an ex-ante unknown partisan component.10 The voter’s payoff is equal to the voter-welfare component, and each party’s payoff depends on both the voter-welfare and partisan components. Alternatives that have a small (respectively, large) partisan component provide both parties (resp., only one party) with a positive payoff and, hence, are said to be bipartisan (resp., divisive). In Section 7, I discuss the key assumptions of the benchmark model and explore some extensions, which are formally analyzed in the Supplementary Material. The first extension allows for an arbitrary (and stochastic) number of alternatives to arrive in each period and, in addition, allows for an arbitrary but finite time horizon. The second extension relaxes the strategic sophistication of the voter by replacing the endogenous election outcome with an exogenous transition probability. The model shows that policy bundling can increase gridlock. Policy bundling will sometimes reduce gridlock via two distinct channels. First, by facilitating compromise: parties may agree to pass a bundle of alternatives if the benefits of passing each alternative that they prefer outweigh the cost of passing alternatives that they don’t. Second, by allowing parties to more fully exploit their proposal power: a small divisive alternative that they favor can be bundled with a bipartisan alternative, which could have other wise passed on its own. This allows for more alternatives to pass to the benefit of the proposer and at the expense of the opposition. However, policy bundling also creates an incentive for parties to gridlock bipartisan alternatives. Preventing a bipartisan alternative from passing is costly in terms of delay, but it also provides benefits. By vetoing a bipartisan alternative, a party may benefit in the future since (if elected) they will be able to bundle this same alternative with future alternatives that are divisive and otherwise would not pass. That is, the vetoed alternative can be used as leverage in future negotiations. This leverage incentive exists only because of policy bundling and can induce gridlock; I call gridlock of this form leverage-based gridlock.11 Leverage-based gridlock differs from gridlock that arises when an alternative is divisive, which I call preference-based gridlock. Preference-based gridlock coincides with the notion of a ‘‘gridlock interval” (Krehbiel, 1996, 1998). Leverage-based gridlock will occur more frequently when alternatives are more partisan or unpassed alternatives are more likely to be available in the future. When a party engages in leveragebased gridlock there are two costs. First, the unpassed alternative may be unavailable in the future, in which case the party forgoes the payoff associated with the alternative and there is no leverage benefit. Second, there is a cost of delay because future payoffs are discounted. As a bipartisan alternative becomes more partisan, one party’s payoff from passing the alternative decreases and the other party’s payoff increases. This leads the party with lower (but positive) payoff to have a greater incentive to engage in leveragebased gridlock: they face a smaller cost of delay and, via the delayed alternative, will have more leverage to pass divisive alternatives in the future. Similarly, leverage-based gridlock will be more frequent when unpassed alternatives are more likely to be available in the future (e.g., during times of economic stability) because the cost of delay of reduced. Contrary to standard models of bargaining, I show that an increase in polarization can have a non-monotonic effect on the frequency of gridlock. I measure polarization by the likelihood that each period’s alternative has a large partisan component. Hence, polarization increases the likelihood that an alternative is divisive. However, in equilibrium, polarization also affects the leverage incentive of parties. When polarization is low or high, there is no leverage incentive and bipartisan alternatives are never gridlocked but, when polarization is intermediate, gridlock may reach its highest levels. Leverage-based gridlock is accentuated by the voter being forward-looking. In the model, the voter rationally elects the party that is likely to pass the most alternatives. This behavior is intuitive—voters want to elect politicians that ‘‘get stuff done” (Rivlin, 2018). However, this also implies that, conditional on an alternative failing to pass, one party will be electorally advantaged and elected with certainty. The forward-looking behavior of the voter leads this electorally-advantaged party to have a greater incentive to engage in leverage-based gridlock.12 With a weaker (or no) electoral advantage, vetoing an alternative comes with additional risk: the other party may be elected and the unpassed alternative used as leverage against the vetoing party. This reduces—but does not eliminate—the incentive for parties to engage in leveraged-based gridlock. Thus, the voter’s desire to elect a politician that will be productive in the future can, perhaps counterintuitively, encourage politicians to gridlock present-day policies. In equilibrium, gridlock occurs in excess. From the voter’s perspective, gridlock is costly but also entails benefits since, if the unpassed alternative is available, more alternatives will pass in the future. When the benefits outweigh the costs, gridlock is said to be efficient. In equilibrium, gridlock always occurs if it is efficient but may also occur when it is inefficient. This state of excess gridlock is driven by the voter’s inability to commit. With commitment power, it is optimal for the voter to follow a ‘‘blame-game” strategy, whereby they never elect a vetoing party; this strategy allows the voter to completely eradicate inefficient leverage-based gridlock. My results have a number of empirical implications. Returning to immigration reform, my results provide insights into why the DREAM Act (and DACA) may be vulnerable to gridlock. Although the policies enjoy bipartisan support, they provide much greater benefit to the Democratic party and, hence, by delaying such policies, Republicans obtain a large amount of leverage to pass divisive issues in the future. My model also suggests that this leverage based gridlock benefits the Republicans electorally since it allows them to credibly claim that, if elected, they’ll be able to pass more policies in the future compared to the Democrats. More generally, the phenomenon of leverage-based gridlock is empirically relevant because it suggests that traditional measures of ideology based on roll-call voting records (e.g., Poole and Rosenthal (1985)’s NOMINATE methods) will overstate ideological disagreement between parties. Furthermore, my model predicts that gridlock will be more frequent during periods of economic stability. These two predictions combine to provide a novel explanation for why measures of polarization have increased rapidly since the 1980s (Barber and McCarty, 2015). The 1980s marked the beginning of the Great Moderation, a period of economic stability, leading to more frequent leverage-based gridlock and, hence, overstated measures of polarization. This may also explain why within-Congress polarization compared to that within the public is more pronounced (Hetherington, 2001). Finally, my results suggest that direct measures of gridlock can also fail to capture ideological disagreement since gridlock and polarization have a non-monotonic relationship.